JOURNAL OF VIROLOGY,0022-538X/01/$04.0010 DOI: 10.1128/JVI.75.2.844–849.2001

Jan. 2001, p. 844–849 Vol. 75, No. 2

Copyright © 2001, American Society for Microbiology. All Rights Reserved.

Sialic Acid Binding Activity of Transmissible GastroenteritisCoronavirus Affects Sedimentation Behavior of Virions and

Solubilized GlycoproteinsCHRISTINE KREMPL1 AND GEORG HERRLER1,2*

Institut fur Virologie, Philipps-Universitat Marburg, 35037 Marburg,1 and Institut fur Virologie, TierarztlicheHochschule Hannover, 30559 Hannover,2 Germany

Received 28 June 2000/Accepted 19 October 2000

The sedimentation behavior of transmissible gastroenteritis coronavirus (TGEV) was analyzed. Upon su-crose gradient centrifugation, the major virus band was found at a density of 1.20 to 1.22 g/cm3. This highdensity was observed only when TGEV with a functional sialic acid binding activity was analyzed. Mutants ofTGEV that lacked sialic acid binding activity due to a point mutation in the sialic acid binding site of the Sprotein were mainly recovered at a lower-density position on the sucrose gradient (1.18 to 1.19 g/cm3).Neuraminidase treatment of purified virions resulted in a shift of the sedimentation value from the higher tothe lower density. These results suggest that binding of sialoglycoproteins to the virion surface is responsiblefor the sedimentation behavior of TGEV. When purified virions were treated with octylglucoside to solubilizeviral glycoproteins, ultracentrifugation resulted in sedimentation of the S protein of TGEV. However, whenneuraminidase-treated virions or mutants with a defective sialic acid binding activity were analyzed, the Sprotein remained in the supernatant rather than in the pellet fraction. These results indicate that theinteraction of the surface protein S with sialoglycoconjugates is maintained after solubilization of this viralglycoprotein by detergent treatment.

Transmissible gastroenteritis coronavirus (TGEV) is a pro-totype enteropathogenic coronavirus that causes diarrhea inpigs of all ages. While adult animals usually recover, newbornpiglets generally die from the intestinal infection (14). TGEVis an enveloped virus with three proteins inserted into the viralmembrane: S (220 kDa), M (29 to 36 kDa), and E (10 kDa), aminor protein. The S protein plays a key role in the initial stageof infection. It mediates binding of the virus to the cell surfaceand the subsequent fusion between the viral and cellular mem-branes. Two binding activities have been assigned to the Sprotein. Binding to porcine aminopeptidase N, a cellular re-ceptor for TGEV, is a prerequisite for infection of cells (5). Asecond binding activity enables TGEV to recognize sialic acidresidues and attach to sialoglycoconjugates (18). As a conse-quence of the latter binding activity, TGEV can agglutinateerythrocytes (11, 12). The binding site for aminopeptidase Nand the binding site for sialic acid are located on differentportions of the S protein (18). Recent studies with mutants ofTGEV indicated that a short stretch of amino acids (145 to209) is important for the recognition of sialic acids (8). Someof the mutants had been selected for resistance to a monoclo-nal antibody. Interestingly, the point mutations that were re-sponsible for the lack of antibody reactivity also resulted in theconcomitant loss of both hemagglutinating activity and entero-pathogenicity (8). These results indicated not only that therespective amino acids are located at or close to the sialic acidbinding site but also that the sialic acid binding activity iscorrelated with the enteropathogenicity of TGEV. Other fac-

tors may also be required to render TGEV enteropathogenic,but they have not been identified in terms of a molecularinteraction. The importance of the sialic acid binding activityfor enteropathogenicity is supported by data reported for por-cine respiratory coronavirus (PRCoV), which is closely relatedto TGEV. This virus replicates with high efficiency in the re-spiratory tract but with very low efficiency in the gut (4). Likethe mutants mentioned above, PRCoV has no hemagglutinat-ing activity (18). In the case of PRCoV, the lack of a sialic acidbinding activity is explained by a large deletion in the S genethat results in a truncated spike protein (15, 16). The pointmutations that resulted in the loss of hemagglutinating activityand enteropathogenicity are located in that portion of the Sprotein that is present in the TGEV S protein but absent fromthe PRCoV S protein.

Here we report that the ability of TGEV to attach to sialo-glycoconjugates affects the physical state of virus particles.TGEV with a functional sialic acid binding activity was recov-ered from sucrose gradients at higher densities than PRCoV ormutants of TGEV with a defect in the sialic acid binding site.The difference in sedimentation behavior was also observedwith solubilized S protein. It was abolished after neuramini-dase treatment of virions, suggesting that bound sialoglycop-roteins are responsible for the different sedimentation charac-teristics.

(Part of this work was done by C.K. in partial fullfilment ofthe requirements for the Dr. rerum physiologicarum degree atPhilipps-Universitat Marburg, Marburg, Germany.)

MATERIALS AND METHODS

Virus. The Purdue strain of TGEV (PUR46-MAD) was used throughout thisstudy. Stock virus was propagated in swine testicular (ST) cells. After incubation

* Corresponding author. Mailing address: Institut fur Virologie,Tierarztliche Hochschule Hannover, Bunteweg 17, 30559 Hannover,Germany. Phone: 49 (0) 511-28-8857. Fax: 49 (0) 511-28-8898. E-mail:herrler@viro.tiho-hannover.de.

844

for 20 to 24 h at 37°C, the supernatant was harvested, clarified by low-speedcentrifugation, and after addition of 1% fetal calf serum (FCS), stored at 280°C.

Cells. ST cells were grown in Dulbecco’s modified Eagle medium supple-mented with 10% FCS.

Sucrose gradient centrifugation of TGEV and TGEV mutants. Confluentmonolayers of ST cells were infected with stock virus (PUR46 or HAD mutants)at a multiplicity of infection of 0.1. After incubation for 2 days at 37°C in mediumwithout FCS, the supernatant was clarified by centrifugation for 10 min at1,400 3 g. Virus was sedimented by ultracentrifugation at 140,000 3 g for 1 h.The virus pellet was resuspended in phosphate-buffered saline (PBS) and—ifindicated—treated with neuraminidase as described below. For purification, thevirus was layered onto a 20 to 60% (wt/wt) sucrose gradient (in PBS) andcentrifuged for 4 h at 150,000 3 g. In some experiments, the gradient washarvested from the bottom of the tube and 10 fractions were collected. Thedensity of the sucrose in each fraction was determined using a refractometer.Each fraction was also analyzed for hemagglutinating activity and for the pres-ence of viral proteins by sodium dodecyl sulfate-polyacrylamide gel electrophore-sis (SDS-PAGE). In other experiments, virus bands were harvested and, follow-ing dilution with PBS, were sedimented by centrifugation for 1 h at 150,000 3 g.Purified virus was analyzed by SDS-PAGE or stored at 220°C.

Sedimentation of virions. TGEV and the HAD7 mutant were grown in STcells. After clarification by low-speed centrifugation, aliquots of the supernatantwere centrifuged for 1 h at 0 (control), 25,000, 50,000, 75,000, or 100,000 3 g. Inorder to pellet virions that had not been sedimented in the first centrifugationstep, the supernatants were subjected to centrifugation for 1 h at 150,000 3 g.The pellets of each sample were analyzed by SDS-PAGE. The presence of viruswas determined by Western blotting using a monoclonal antibody (6A.C3) di-rected to the S protein (3).

Neuraminidase treatment of virus. Virus that had been sedimented was re-suspended in 200 ml of PBS per tube and treated with 50 mU of Vibrio choleraeneuraminidase per ml for 30 min at 37°C. Treated virus was purified by sucrosegradient centrifugation as described above.

Octylglucoside treatment of virions. Purified virions were treated with octyl-glucoside at a final concentration of 1 or 5%. The sample was incubated for 1 hon ice. For sedimentation of the nucleocapsid, the lysate was centrifuged in aTLA 45 rotor for 1 h at 100,000 3 g at 4°C. The supernatants and the pellet

fractions were analyzed for the presence of the S protein by Western blottingusing a monoclonal antibody as described above.

Electron microscopy. For negative staining, samples were applied to coppergrids, stained with phosphotungstate, and examined with a Zeiss electron micro-scope 110.

Western blotting. Viral proteins were separated by SDS-PAGE under nonre-ducing conditions and blotted to nitrocellulose using a semidry Western blotmethod, as described by Schultze et al. (17). After blocking of nonspecificbinding sites by incubation with 10% nonfat dry milk in PBS, the immobilizedTGEV S protein was incubated with different monoclonal antibodies. Boundantibodies were detected with biotinylated sheep anti-mouse immunoglobulin G(Amersham) and a streptavidin-biotinylated horseradish peroxidase complex(Amersham).

For detection of proteins in polyacrylamide gels, a silver stain (Bio-Rad) wasused according to the instructions of the manufacturer.

RESULTS

Sucrose gradient centrifugation of TGEV. The sialic acidbinding activity of TGEV may result in the binding of sialo-glycoconjugates to the viral surface. As glyoproteins usuallycontain more than one oligosaccharide, they have the potentialto establish a multivalent interaction with two or more S pro-teins. Therefore, a single sialoglycoconjugate may even attachto two virions. We wondered whether such interactions affectthe sedimentation behavior of TGEV and analyzed a purifiedvirus preparation by sucrose gradient centrifugation. As shownin Fig. 1, the S protein was mainly found in fractions 2 and 3(upper left panel), corresponding to sucrose concentrations ofabout 45 to 49% (lower left panel) or a density of 1.20 to 1.22g/cm3. When the sample was treated with sialidase prior toultracentrifugation to remove bound sialoglycoconjugates fromthe virion surface, the majority of the virus particles were found in

FIG. 1. Sucrose gradient centrifugation of untreated (TGEV) and neuraminidase-treated (TGEV-NA) TGEV. The gradient was fractionatedfrom the bottom. Aliquots of each fraction were used to measure the sucrose concentration (lower panels) and to analyze by SDS-PAGE for thepresence of virus (upper panels). The bands were stained with Coomassie brilliant blue. Only the portion of the gel containing the S protein isshown.

VOL. 75, 2001 SEDIMENTATION CHARACTERISTICS OF TGEV 845

fractions 4 and 5 (upper right panel), corresponding to sucroseconcentrations of about 40 to 43% (lower right panel) or a densityof 1.18 to 1.19 g/cm3. This value is close to the value determinedfor coronaviruses (1.15 to 1.18 g/cm3) by equilibrium sedimenta-tion (2). This result indicates that sialic acid binding activity canaffect the sedimentation behavior of TGEV.

Upon ultracentrifugation of TGEV through sucrose gradi-ents, we usually observed two bands. In light of the aboveresults, these bands may represent populations of virions, thesedimentation of which has been affected or unaffected bysialic acid binding activity. For clarification, untreated andneuraminidase-treated preparations of TGEV were centri-fuged through a 20 to 60% sucrose gradient, and the upper andlower bands were harvested and analyzed by SDS-PAGE. Asshown in Fig. 2A, the majority of virions in the untreatedfraction are present in the lower band (designated 2). This isespecially evident in the S protein. Neuraminidase-treated viri-ons are mainly present in the upper band (designated 1). Todemonstrate that the positions of the two bands are actuallydetermined by sialic acid binding activity, we also analyzedmutants of TGEV (Fig. 2B). Mutant HAD7 had been selectedfor its inability to bind to erythrocytes and was shown to lacksialic acid binding activity (9). Virions of this mutant weremainly found in the upper band irrespective of neuraminidasetreatment. The same picture was observed with mutant m6.This mutant had been selected for its resistance to a monoclo-nal antibody (1) and was shown to lack hemagglutinating ac-

tivity (8). Mutant m9 was selected for resistance to the samemonoclonal antibody but has retained its hemagglutinatingactivity (1, 8). This mutant resembled the parental TGEV; i.e.,untreated virions were mainly present in the lower band,whereas neuraminidase-treated virions were mainly detectedin the upper band.

Sedimentation of TGEV. In order to find out whether thedifference between hemagglutinating and nonhemagglutinat-ing TGEV is detectable only at the high centrifugation forcesapplied during sucrose gradient centrifugation, we analyzedthe sedimentation behavior under different conditions. Ali-quots of the parental TGEV and the HAD7 mutant weresedimented at 25,000, 50,000, 75,000, or 100,000 3 g. Thepellets were each resuspended in PBS, and both pellet andsupernatant fractions were sedimented at 150,000 3 g. Thepellets were subjected to SDS-PAGE and probed with a mono-clonal antibody for the presence of S protein. As shown in Fig.3, at a centrifugal force as low as 25,000 3 g the majority of thevirions (PUR46) were found in the pellet fraction; no and onlytrace amounts of S protein were detectable in the supernatantafter centrifugation at 75,000 and 100,000 3 g, respectively.With the HAD7 mutants, on the other hand, even after cen-trifugation at 100,000 3 g about half of the virions were stillpresent in the supernatant fraction. This result confirms thatthe presence or absence of a functional sialic acid bindingactivity in TGEV may affect the sedimentation behavior ofvirions.

FIG. 2. Sucrose gradient centrifugation of untreated and neuraminidase-treated virions. The upper right panel shows a schematic drawing ofthe two bands harvested. The upper left panel shows the S, N, and M protein profiles of the upper (1) and lower (2) bands of untreated (TGEV)and neuraminidase-treated (TGEV-NA) TGEV. The lower panel shows the S and N protein profiles of the upper (1) and lower (2) bands of them9, m6, and HAD7 mutants. Lanes representing untreated virions (2) and neuraminidase-treated virions (1) are indicated. The left lane (M)shows molecular mass markers. The proteins were stained with Coomassie brilliant blue.

846 KREMPL AND HERRLER J. VIROL.

Lysis of virions by octylglucoside. In order to determinewhether the sialic acid binding activity also affects sedimenta-tion of the glycoprotein S, virions were solubilized by octylglu-coside. The nucleocapsid was sedimented by centrifugation at100,000 3 g. Under these conditions, viral glycoproteins areexpected in the supernatant fraction. In order to verify theeffectiveness of the solubilization, the supernatant and thepellet fractions were analyzed for the presence of S protein. Asshown in Fig. 4, neuraminidase-treated TGE virions showedthe characteristics expected for an enveloped virus. In theabsence of octylglucoside, the S protein was present in thepellet fraction. After treatment with 1% octylglucoside, morethan 50% of the S protein was detected in the supernatant.With virions treated with 5% octylglucoside, all of the S pro-tein was found in the supernatant fraction. A different picturewas obtained with purified TGEV that had not been pretreatedwith neuraminidase. Even at the highest concentration of oc-tylglucoside, no S protein was detectable in the supernatantfraction. To confirm that the difference was due to the sialicacid binding activity, PRCoV, a nonhemagglutinating variantof TGEV, and several mutants of TGEV were analyzed in thesame way. As shown in Fig. 5, the viruses that lacked a hem-agglutinating activity resembled each other; i.e., after treat-ment with 1% octylglucoside, the majority of the S protein wasdetected in the supernatant fraction. In contrast, the hemag-glutinating m9 mutant behaved after octylglucoside treatment

like the parental virus, with the S protein being present mainlyin the pellet fraction.

The appearance of the S protein in the pellet fraction mightsuggest that the virions had resisted detergent treatment. How-ever, no evidence of the presence of intact virions was obtainedby analysis with electron microscopy. Rather, the pellet frac-tion contained rosette-like structures (Fig. 6), as described forother coronavirus glycoproteins after removal of detergent(17). This result indicates that the interactions with sialoglyco-conjugates that changed the sedimentation behavior of TGEvirions were maintained after detergent lysis of the viral mem-brane and resulted in sedimentation of the S protein despitethe presence of detergent.

DISCUSSION

Our comparative analysis has demonstrated that TGEV mu-tants lacking a sialic acid binding activity have the sedimenta-

FIG. 3. Sedimentation of TGEV and the HAD7 mutant. Purified virions were sedimented for 1 h at 0 (lane 1), 25,000 (lane 2), 50,000 (lane3), 75,000 (lane 4), and 100,000 3 g (lane 5). Virions remaining in the supernatant were sedimented by centrifugation for 1 h at 150,000 3 g. Thepellet fractions of the first centrifugation (P) and the second centrifugation (S) were analyzed by Western blotting for the presence of S protein.

FIG. 4. Sedimentation of the S protein after solubilization withoctylglucoside (OG). Untreated (2VCNA) and neuraminidase-treated (1VCNA) virions were solubilized with 0, 1, or 5% octylglu-coside and sedimented by ultracentrifugation as described in Materialsand Methods. Aliquots of the pellet (P) and the supernatant (S) frac-tions were analyzed by Western blotting for the presence of S protein.

FIG. 5. Sedimentation of the S proteins of TGEV, PRCoV, andmutants of TGEV (HAD2, HAD3, HAD4, HAD7, m9, and m6) aftersolubilization with octylglucoside (OG). The purified virus prepara-tions were not pretreated with neuraminidase. Virions were incubatedwith 0 or 1% octylglucoside and sedimented by ultracentrifugation asdescribed in Materials and Methods. Aliquots of the pellet (P) and thesupernatant (S) fractions were analyzed by Western blotting for thepresence of S protein. HA, hemagglutinating.

VOL. 75, 2001 SEDIMENTATION CHARACTERISTICS OF TGEV 847

tion characteristics reported for coronaviruses (2). By contrast,the parental virus banded at a higher density of the sucrosegradient. As neuraminidase treatment of virions abolished thedifference in sedimentation behavior, we conclude that bindingof sialoglycoproteins to the virion surface is responsible for theincreased density of untreated TGEV. The binding of sialy-lated glycoconjugates to TGEV has been suggested before,because virions harvested late in infection are not able toagglutinate erythrocytes (18). The hemagglutinating activitycan be restored by neuraminidase treatment. The enzymetreatment obviously removes sialic acid residues that occupythe sialic acid binding site, allowing the S protein to interactwith sialic acid residues on the erythrocyte. As the S protein ofTGEV is highly glycosylated, one might assume that the sialicacid binding activity of this viral spike protein might interactwith a sialic acid residue of another S protein, either on thesame virion or on another virus particle. However, the avail-able evidence indicates that mainly sialoglycoconjugates de-rived from the cell surface are bound to the viral surface.TGEV released into the cell supernatant early in infection isable to agglutinate erythrocytes. The hemagglutinating activityis lost late in infection and can be restored by neuraminidasetreatment (18). Rather than treating the virions with enzyme,the hemagglutinating activity can also be restored by treatingthe confluent cell monolayer with neuraminidase prior to in-fection (8). This method does not affect the sialylation of the Sprotein; nevertheless, it is almost as efficient in obtaining hem-agglutinating TGEV as enzyme treatment of virions. This find-ing suggests that sialoglycoconjugates present on the surface ofthe confluent cell monolayer are the major ligands that bind tothe S protein and prevent it from interacting with erythrocytes.These ligands are presumably also responsible for the sedi-mentation behavior of TGEV described above.

The binding of sialoglycoconjugates to the virion surfacedoes not affect virus infectivity. Whether or not TGEV istreated with neuraminidase, the infectivity titer is the same(18). This may be explained by the findings that sialic acidbinding activity is not required for infection of cells and thatthe binding sites for aminopeptidase and for sialic acid arelocated at different positions on the S protein. Amino acidsinvolved in the binding to aminopeptidase have been identifiedbetween positions 506 and 728 of the mature polypeptide chain(6). Binding to sialic acid was found to require amino acidsbetween positions 145 and 209 (8, 9). Though the three-dimen-sional structure of the S protein has not yet been elucidated, itappears likely that the two binding sites are located at somedistance from each other. Moreover, with virions containingsurface-bound sialoglycoconjugates, not every spike protein onthe viral surface, but only a fraction of them, is expected toactually interact with a sialylated ligand. This would be suffi-cient to prevent the virus from agglutinating erythrocytes, be-cause the same type of binding site is involved and because ahemagglutination pattern requires the interaction of a virusparticle with two erythrocytes. On the other hand, for theinitiation of an infection, a virion has to attach to only a singlecell. Even if several spike proteins contain sialylated ligands,there will be a sufficient number of S proteins left to interactwith aminopeptidase N on the cell surface.

Influenza viruses or paramyxoviruses that use sialic acid as areceptor determinant for infection of cells contain a neuramin-idase. This “receptor-destroying” enzyme is required for effi-cient spread of virions from the infected cell to other cells.With neuraminidase-deficient mutants or in the presence ofneuraminidase inhibitors, aggregates of virions accumulate onthe cell surface (10, 13). No such occurrences have been re-ported for TGEV, though the virus does not have a compara-

FIG. 6. Electron micrograph of the pellet fraction obtained after centrifugation of octylglucoside-treated virions. Magnification, 3150,000. Bar,100 nm.

848 KREMPL AND HERRLER J. VIROL.

ble enzyme. This may be related to the different mode of virusmaturation. Influenza viruses and paramyxoviruses are re-leased from the cell by a budding process at the cell surface.Coronaviruses, on the other hand, mature on intracellularmembranes, and the virions are transported within membranevesicles to the cell surface (19). Fusion of the transport vesicleswith the plasma membrane results in release of virions into thecell supernatant. This mode of maturation may be more favor-able with the interaction of virions with sialoglycoconjugatesthat are only loosely attached to cells than with membrane-anchored glycoproteins or glycolipids. Furthermore, S proteinsnot incorporated into virions are transported to the plasmamembrane. They may interact with membrane-anchored sialo-glycoconjugates and prevent them from binding to virions.Furthermore, the sialic acid binding activity of influenza virusmay be stronger than that of TGEV and therefore may requirea receptor-destroying enzyme. Recently, mutants of influenzaviruses that lacked neuraminidase activity but could neverthe-less undergo several replication cycles in cell cultures, eggs,and mice have been described (7). These mutants compen-sated for the loss of sialidase by a reduced affinity for cellularreceptors.

The role of the sialic acid binding activity is not yet known.We reported recently that TGEV virions containing surface-bound sialoglycoconjugates have increased resistance to theaction of detergents (9). Neuraminidase-treated TGEV or mu-tants deficient in sialic acid binding activity were completelyinactivated by 0.6% octylglucoside. On the other hand, theinfectivity of untreated TGEV was only partially inactivated atthis concentration of detergent. Therefore, surface-bound sia-loglycoconjugates may be beneficial during passage throughthe gastrointestinal tract and may protect the virus from thedetrimental action of detergent-like substances, such as bilesalts. Another possibility is that the sialic acid binding activitycontributes to the binding of TGEV to epithelial cells of theintestine. While interaction with aminopeptidase N is sufficientfor infection of cultured cells, the initiation of infection in theintestinal environment may require additional receptors to ac-complish efficient binding to enterocytes. Attempts to identifysialoglycoconjugates on the surface of the intestinal epitheliumthat may serve as attachment sites for TGEV are in progress.

ACKNOWLEDGMENTS

The technical assistance of Anja Heiner is gratefully acknowledged.We are grateful to Luis Enjuanes for providing monoclonal antibody6A.C3 and to Hubert Laude for providing site D mutants.

Financial support was provided by Deutsche Forschungsgemein-schaft (He1168/2-3 and SFB 280).

REFERENCES

1. Bernard, S., and H. Laude. 1995. Site-specific alteration of transmissiblegastroenteritis virus spike protein results in markedly reduced pathogenicity.J. Gen. Virol. 76:2235–2241.

2. Cavanagh, D., D. A. Brian, M. A. Brinton, L. Enjuanes, K. V. Holmes, M. C.Horzinek, M. M. C. Lai, H. Laude, P. G. W. Plagemann, S. G. Siddell, W.Span, F. Taguchi, and P. J. Talbot. 1994. Revision of the taxonomy of theCoronavirus, Torovirus and Arterivirus genera. Arch. Virol. 135:227–237.

3. Correa, I., G. Jimenez, C. Sune, M. J. Bullido, and L. Enjuanes. 1988.Antigenic structure of the E2 glycoprotein from transmissible gastroenteritiscoronavirus. Virus Res. 10:77–93.

4. Cox, E., M. B. Pensaert, P. Callebaut, and K. van Deun. 1990. Intestinalreplication of a porcine respiratory coronavirus closely related antigenicallyto the enteric transmissible gastroenteritis. Vet. Microbiol. 23:237–243.

5. Delmas, B., J. Gelfi, R. L’Haridon, L. K. Vogel, H. Sjostrom, O. Noren, andH. Laude. 1992. Aminopeptidase N is a major receptor for the entero-pathogenic coronavirus TGEV. Nature 357:417–420.

6. Godet, M., J. Grosclaude, B. Delmas, and H. Laude. 1994. Major receptor-binding and neutralization determinants are located within the same domainof the transmissible gastroenteritis virus (coronavirus) spike protein. J. Virol.68:8008–8016.

7. Hughes, M. T., M. Matrosovich, M. E. Rodgers, M. McGregor, and Y.Kawaoka. 2000. Influenza A viruses lacking sialidase activity can undergomultiple cycles of replication in cell culture, eggs, or mice. J. Virol. 74:5206–5212.

8. Krempl, C., B. Schultze, H. Laude, and G. Herrler. 1997. Point mutations inthe S protein connect the sialic acid binding activity with the enteropatho-genicity of transmissible gastroenteritis coronavirus. J. Virol. 71:3285–3287.

9. Krempl, C., M. L. Ballesteros, G. Zimmer, L. Enjuanes, H.-D. Klenk, and G.Herrler. 2000. Characterization of the sialic acid binding activity of trans-missible gastroenteritis coronavirus by analysis of haemagglutination-defi-cient mutants. J. Gen. Virol. 81:489–496.

10. Liu, C., M. C. Eichelberger, R. W. Compans, and G. M. Air. 1995. Influenzatype A virus neuraminidase does not play a role in viral entry, replication,assembly, or budding. J. Virol. 69:1099–1106.

11. Noda, M., F. Koide, M. Asagi, and Y. Inaba. 1988. Physicochemical proper-ties of transmissible gastroenteritis virus hemagglutinin. Arch. Virol. 99:163–172.

12. Noda, M., H. Yamashita, F. Koide, K. Kadoi, T. Omori, M. Asagi, and Y.Inaba. 1987. Hemagglutination with transmissible gastroenteritis virus. Arch.Virol. 96:109–115.

13. Palese, P., and R. W. Compans. 1976. Inhibition of influenza virus replicationin tissue culture by 2-deoxy-2,3-dehydro-N-trifluoroacetylneuraminic acid(FANA): mechanism of action. J. Gen. Virol. 33:159–163.

14. Pensaert, M., P. Callebaut, and E. Cox. 1993. Enteric coronaviruses ofanimals, p. 627–696. In A. Z. Kapikian (ed.), Viral infections of the gastro-intestinal tract. Marcel Dekker, New York, N.Y.

15. Rasschaert, D., M. Duarte, and H. Laude. 1990. Porcine respiratory coro-navirus differs from transmissible gastroenteritis virus by a few genomicdeletions. J. Gen. Virol. 71:2599–2607.

16. Sanchez, C. M., F. Gebauer, C. Sune, A. Mendez, J. Dopazo, and L. En-juanes. 1992. Genetic evolution and tropism of transmissible gastroenteritiscoronaviruses. Virology 190:92–105.

17. Schultze, B., H.-J. Gross, R. Brossmer, and G. Herrler. 1991. The S proteinof bovine coronavirus is a hemagglutinin recognizing 9-O-acetylated sialicacid as a receptor determinant. J. Virol. 65:6232–6237.

18. Schultze, B., C. Krempl, M. L. Ballesteros, L. Shaw, R. Schauer, L. En-juanes, and G. Herrler. 1996. Transmissible gastroenteritis coronavirus, butnot the related porcine respiratory coronavirus, has a sialic acid (N-glyco-lylneuraminic acid) binding activity. J. Virol. 70:5634–5637.

19. Tooze, J., S. Tooze, and G. Warren. 1984. Replication of coronavirus MHV-A59 in sac2 cells: determination of the first site of budding of progenyvirions. Eur. J. Cell Biol. 33:281–293.

VOL. 75, 2001 SEDIMENTATION CHARACTERISTICS OF TGEV 849